tbl01 <-table(ds$Country)## Table of user language by country:tbl01
Gambia Pakistan Tanzania Uganda
237 505 352 448
Display code
## Sample size by country:lp01 <- ds %>%# drop_na(Country) %>%lollipop_chart(x = Country,line_color ="black",point_color ="black")+labs(y ="Frequency",x ="",title ="Sample size by country")+theme_bw()lp01
Variable: Age
Display code
summary(ds$age)
Min. 1st Qu. Median Mean 3rd Qu. Max. NA's
18 26 33 37 45 92 29
Display code
ds %>%drop_na(age)%>%ggplot(aes(x = age))+geom_histogram(color ="black",fill ="gray",bins =50)+geom_textvline(label ="Mean = 37.00", xintercept =37.00, vjust =1.1, lwd =1.05, linetype =2)+labs(x ="Age", y ="Frequency", title ="Age distribution (full sample)")+theme_bw()
Display code
ds %>%drop_na(age)%>%ggplot(aes(x = age))+geom_histogram(color ="black",fill ="gray",bins =50)+labs(x ="Age", y ="Frequency", title ="Age distribution by country")+facet_wrap(~Country, nrow =2)+theme_bw()
44
4
8
4m 4
1
7
7
Advance Diploma
1
Advance Diploma in Business Studies
1
Advance Diploma in Management
3
Advance Diploma in Nursing
1
Advance Diploma in Tourism Management
1
Advance level
1
B.S.C
1
Bachelor
104
Bachelor Degree
2
Bachelor in Business Administration
1
Bachelor's Degree
3
Bachelor's Degree in Political Science
1
Bachelors (ongoing)
1
Bachelors Degree
1
Bachelors in Education
1
Bachelors of Edu.
1
Bsc
1
BSc
3
BSC
5
BSc (Arabic Education)
1
BSc (Hons) Management
1
BSc in Accounting
1
BSc in Economic
1
BSc in Economica
1
BSc in Education
1
BSc in Human Resource
1
BSc in Management
1
BSc in Political Science
1
BsSc of Accounting
1
Certificate
3
Certificate in International Relation
1
Certificate in Management Studies
1
Certificate Level
1
Certificate on Animal Health Product
1
Certtificate
1
Chartered Banker
1
Cheti
6
Chuo
6
Chuo Ckikuu
1
Chuo kikuu
11
Chuo Kikuu
3
Claiasa nne
1
College
9
College (HTC)
1
College (HWD)
1
College Certificate
1
currently in s4
1
Darasa 7
8
Darasa 8
1
Darasa la 7
1
Darasa la Saba
1
Darsa la 7
2
Darsa La 7
3
Darsa la saba
1
Degree
15
Degree in Islamic Studies
1
Degree-Shahada
1
Didato channe
1
Digri/Shahada
1
Diploma
22
Diploma 2 Business Administration
1
Diploma in Community Policing
1
Diploma in ICT
2
Diploma in Management
3
Diploma in Theology
1
Diploma in Travel and Tourism
1
Diplomat
1
Elimu Msingi
1
Form 3 (Old education System)
2
Form 3 (Old Education System)
1
Form 4
15
Form 5
1
Form 5 (Old education System)
2
Form 5 (Old Education System)
1
Form 6
3
Form Five
2
Form Five 5
2
Form Four
73
Form Four 4
3
Form Four 5
1
Form IV
1
Form One 1
1
Form Six
14
Form Three
1
Form Two
6
Form Two 2
1
Gambia College (PTC)
1
GCE Level(WASSCE)
1
Grade 8 ( Upper Basic)
1
Graduate Diploma
1
High School
4
High School (WASSCE)
1
High School Certificate
2
High School Level
1
High School-Islamic Education
1
Higher Diploma
1
Higher Diploma in Education (HDC)
1
Higher Teacher Certificate (HTC)
7
Higher Teacher's Certificate (HTC)
5
Hotel School
1
HTC
2
HTC Primary
1
HTC/HNC
1
Illiterate
10
Informal Education (Madrasa)
1
Intermediate
85
Isachetovs Degree
1
Islamic School
1
Islamic Traditional School
1
IV
1
Junior Sec School
1
Junior Secondary
1
Kidato 4
3
Kidato cha 1
1
Kidato cha 4
1
Kidato cha 6
2
Kidato cha nne
2
Kidato Cha seta
1
Kidato Channe
1
Kidato chaune
1
La 4
3
La 7
7
LA 7
8
La nne
4
La nne (4)
1
La nne mkoloni
1
La saba
17
La saba 7
8
Lanne Mkoloni
2
Lasaba
1
LLB
1
Master
161
Masters Degree
2
Masters in Economics
1
Masters in Sociology
1
Matriculate
34
Memorized the Quran (Informal Education)
1
Middle
53
MPhil
56
MSC
1
Msingi
3
None
50
O Level
3
Ordinary Levels
1
PhD
3
Post Graduate
2
Post Graduate Diploma
1
Primary
213
Primary School
3
Primary Teacher Certificate (PTC)
2
Primary Teacher's Certificate (PTC)
1
PTC
1
Quranic High School Graduate
1
Quranic School
1
Seconary 4
1
Secondary
143
Secondary Four
1
Secondary level
1
Secondary School
6
Sekonorw
1
Senior School Graduation
1
Senior Secondary
2
Senior Secondary School
16
Senior Secondary(WASSCE Certificate)
1
Shadada
3
Shahada
5
Shahada yai
1
Shule Msiry
1
Shule ya msingi
1
Sokondal
1
StaShahada
2
Std VII
3
STD VII
8
Tertiary
33
Tertiary Education
10
Tertiary level
1
Traditional Quranic Schools
1
Udergraduate
1
Undergraduate Degree in Communications
1
University
15
University Degree
7
University of the Gambia
1
Uzamili
1
VI
1
VII Std
1
Vocational
8
WASSCE
16
WASSCE examination certificate
1
Min. 1st Qu. Median Mean 3rd Qu. Max. NA's
1 3 5 4 6 7 42
Display code
ds %>%drop_na(bbl)%>%ggplot(aes(x = bbl))+geom_histogram(color ="black",fill ="gray",bins =25)+geom_textvline(label ="Mean = 4.00", xintercept =4.00, vjust =1.1, lwd =1.05, linetype =2)+labs(x ="Endorsement of BBL score", y ="Frequency", title ="Endorsement of BBL (full sample)")+theme_bw()
Display code
ds %>%drop_na(bbl, Country)%>%ggplot(aes(x = bbl))+geom_histogram(color ="black",fill ="gray",bins =25)+labs(x ="Endorsement of BBL score", y ="Frequency", title ="Endorsement of BBL by country")+facet_wrap(~Country)+theme_bw()
Display code
ds %>%drop_na(bbl)%>%ggplot(aes(x = bbl,y = Country))+geom_boxplot(fill ="grey")+geom_textvline(label =" ", xintercept =4.00, vjust =1.1, lwd =1.05, linetype =2)+labs(x ="Endorsement of BBL score", y ="Frequency", title ="Endorsement of BBL by country")+#facet_wrap(~Country, nrow = 2)+theme_bw()
Variable: Endorsement of Barrier Crossing Leadership (BCL)
Display code
summary(ds$bcl)
Min. 1st Qu. Median Mean 3rd Qu. Max. NA's
1 5 6 6 6 7 38
Display code
ds %>%drop_na(bcl)%>%ggplot(aes(x = bcl))+geom_histogram(color ="black",fill ="gray",bins =25)+geom_textvline(label ="Mean = 6.00", xintercept =6.00, vjust =1.1, lwd =1.05, linetype =2)+labs(x ="Endorsement of BCL score", y ="Frequency", title ="Endorsement of BCL (full sample)")+theme_bw()
Display code
ds %>%drop_na(bcl, Country)%>%ggplot(aes(x = bcl))+geom_histogram(color ="black",fill ="gray",bins =25)+labs(x ="Endorsement of BCL score", y ="Frequency", title ="Endorsement of BCL by country")+facet_wrap(~Country)+theme_bw()
Display code
ds %>%drop_na(bcl)%>%ggplot(aes(x = bcl,y = Country))+geom_boxplot(fill ="grey")+geom_textvline(label =" ", xintercept =6.00, vjust =1.1, lwd =1.05, linetype =2)+labs(x ="Endorsement of BCL score", y ="Frequency", title ="Endorsement of BCL by country")+#facet_wrap(~Country, nrow = 2)+theme_bw()
ds %>%drop_na(perspective_taking, Country)%>%ggplot(aes(x = perspective_taking))+geom_histogram(color ="black",fill ="gray",bins =25)+labs(x ="Perspective taking score", y ="Frequency", title ="Perspective taking by country")+facet_wrap(~Country)+theme_bw()
Display code
ds %>%drop_na(perspective_taking)%>%ggplot(aes(x = perspective_taking,y = Country))+geom_boxplot(fill ="grey")+geom_textvline(label =" ", xintercept =6.00, vjust =1.1, lwd =1.05, linetype =2)+labs(x ="Perspective taking score", y ="Frequency", title ="Perspective taking by country")+#facet_wrap(~Country, nrow = 2)+theme_bw()
Variable: Perceived history of discrimination
Display code
summary(ds$history_discrimination)
Min. 1st Qu. Median Mean 3rd Qu. Max. NA's
1 2 4 4 6 7 26
Display code
ds %>%drop_na(history_discrimination)%>%ggplot(aes(x = history_discrimination))+geom_histogram(color ="black",fill ="gray",bins =10)+geom_textvline(label ="Mean = 4.00", xintercept =4.00, vjust =1.1, lwd =1.05, linetype =2)+labs(x ="Perceived history of discrimination score", y ="Frequency", title ="Perceived history of discrimination (full sample)")+theme_bw()
Display code
ds %>%drop_na(history_discrimination, Country)%>%ggplot(aes(x = history_discrimination))+geom_histogram(color ="black",fill ="gray",bins =10)+labs(x ="Perceived history of discrimination score", y ="Frequency", title ="Perceived history of discrimination by country")+facet_wrap(~Country)+theme_bw()
Display code
ds %>%drop_na(history_discrimination)%>%ggplot(aes(x = history_discrimination,y = Country))+geom_boxplot(fill ="grey")+geom_textvline(label =" ", xintercept =4.00, vjust =1.1, lwd =1.05, linetype =2)+labs(x ="Perceived history of discrimination score", y ="Frequency", title ="Perceived history of discrimination by country")+#facet_wrap(~Country, nrow = 2)+theme_bw()
Variable: Event: Positive Affect
Display code
summary(ds$Event_positive_affect)
Min. 1st Qu. Median Mean 3rd Qu. Max. NA's
1 1 2 4 6 7 32
ds %>%drop_na(Event_reflection, Country)%>%ggplot(aes(x = Event_reflection))+geom_histogram(color ="black",fill ="gray",bins =15)+labs(x ="Event: Reflection score", y ="Frequency", title ="Event: Reflection by country")+facet_wrap(~Country)+theme_bw()
Display code
ds %>%drop_na(Event_reflection)%>%ggplot(aes(x = Event_reflection,y = Country))+geom_boxplot(fill ="grey")+geom_textvline(label =" ", xintercept =6.00, vjust =1.1, lwd =1.05, linetype =2)+labs(x ="Event: Reflection score", y ="Frequency", title ="Event: Reflection by country")+#facet_wrap(~Country, nrow = 2)+theme_bw()
Variable: Event: Transformative for individual
Display code
summary(ds$Event_transformative_indiv)
Min. 1st Qu. Median Mean 3rd Qu. Max. NA's
1 4 6 5 6 7 49
Display code
ds %>%drop_na(Event_transformative_indiv)%>%ggplot(aes(x = Event_transformative_indiv))+geom_histogram(color ="black",fill ="gray",bins =15)+geom_textvline(label ="Mean = 5.00", xintercept =5.00, vjust =1.1, lwd =1.05, linetype =2)+labs(x ="Event: Transformative for individual score", y ="Frequency", title ="Event: Transformative for individual (full sample)")+theme_bw()
Display code
ds %>%drop_na(Event_transformative_indiv, Country)%>%ggplot(aes(x = Event_transformative_indiv))+geom_histogram(color ="black",fill ="gray",bins =15)+labs(x ="Event: Transformative for individual score", y ="Frequency", title ="Event: Transformative for individual by country")+facet_wrap(~Country)+theme_bw()
Display code
ds %>%drop_na(Event_transformative_indiv)%>%ggplot(aes(x = Event_transformative_indiv,y = Country))+geom_boxplot(fill ="grey")+geom_textvline(label =" ", xintercept =5.00, vjust =1.1, lwd =1.05, linetype =2)+labs(x ="Event: Transformative for individual score", y ="Frequency", title ="Event: Transformative for individual by country")+#facet_wrap(~Country, nrow = 2)+theme_bw()
Variable: Event: Transformative for group
Display code
summary(ds$Event_transformative_group)
Min. 1st Qu. Median Mean 3rd Qu. Max. NA's
1.0 4.0 5.0 4.9 6.0 7.0 28
Display code
ds %>%drop_na(Event_transformative_group)%>%ggplot(aes(x = Event_transformative_group))+geom_histogram(color ="black",fill ="gray",bins =15)+geom_textvline(label ="Mean = 5.00", xintercept =5.00, vjust =1.1, lwd =1.05, linetype =2)+labs(x ="Event: Transformative for group score", y ="Frequency", title ="Event: Transformative for group (full sample)")+theme_bw()
Display code
ds %>%drop_na(Event_transformative_group, Country)%>%ggplot(aes(x = Event_transformative_group))+geom_histogram(color ="black",fill ="gray",bins =15)+labs(x ="Event: Transformative for group score", y ="Frequency", title ="Event: Transformative for group by country")+facet_wrap(~Country)+theme_bw()
Display code
ds %>%drop_na(Event_transformative_group)%>%ggplot(aes(x = Event_transformative_group,y = Country))+geom_boxplot(fill ="grey")+geom_textvline(label =" ", xintercept =5.00, vjust =1.1, lwd =1.05, linetype =2)+labs(x ="Event: Transformative for group score", y ="Frequency", title ="Event: Transformative for group by country")+#facet_wrap(~Country, nrow = 2)+theme_bw()
ds %>%drop_na(Event_imagistic, Country)%>%ggplot(aes(x = Event_imagistic))+geom_histogram(color ="black",fill ="gray",bins =15)+labs(x ="Event: Imagistic score", y ="Frequency", title ="Event: Imagistic by country")+facet_wrap(~Country)+theme_bw()
Display code
ds %>%drop_na(Event_imagistic)%>%ggplot(aes(x = Event_imagistic,y = Country))+geom_boxplot(fill ="grey")+geom_textvline(label =" ", xintercept =35.00, vjust =1.1, lwd =1.05, linetype =2)+labs(x ="Event: Imagistic score", y ="Frequency", title ="Event: Imagistic by country")+#facet_wrap(~Country, nrow = 2)+theme_bw()
Section 3. Outcome: Social perception of Religious Freedom
Unconditional means model
Also called “varying intercept model with no predictors” (Gelman and Hill, 2016, Chapter 12). Allows intercepts to randomly vary across countries:
Display code
ds$SPRF <- ds$Sprf## Varying intercept model with no predictors:m00<-lmer(SPRF ~1+ (1| Country), data = ds)summary(m00)
Linear mixed model fit by REML. t-tests use Satterthwaite's method [
lmerModLmerTest]
Formula: SPRF ~ 1 + (1 | Country)
Data: ds
REML criterion at convergence: 3068
Scaled residuals:
Min 1Q Median 3Q Max
-4.726 -0.572 0.096 0.676 2.641
Random effects:
Groups Name Variance Std.Dev.
Country (Intercept) 0.0257 0.160
Residual 0.4502 0.671
Number of obs: 1497, groups: Country, 4
Fixed effects:
Estimate Std. Error df t value Pr(>|t|)
(Intercept) 5.4297 0.0822 3.0385 66 0.0000067 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Random effects:
Variance for Intercept = 0.03. This is the variance of the means across level 1 categories (countries). Residual = 0.44. Variance explained by level 1 residuals (everything that’s not in level 1).
Display code
tab_model(m00)
SPRF
Predictors
Estimates
CI
p
(Intercept)
5.43
5.27 – 5.59
<0.001
Random Effects
σ2
0.45
τ00Country
0.03
ICC
0.05
N Country
4
Observations
1497
Marginal R2 / Conditional R2
0.000 / 0.054
We can see that ICC = 0.06. Lower ICC = low variance explained across groups. In this case, most of the variability is at individual-level (not group level). There is no significantly different patterns between countries.
Random intercept models
Also called “varying intercept model with individual-level predictors” (Gelman and Hill, 2016, Chapter 12).
Linear mixed model fit by REML. t-tests use Satterthwaite's method [
lmerModLmerTest]
Formula: Perception_religious_freedom ~ IG_Fusion + IG_Identification +
OG_Bonds + Empathic_concern + Perspective_taking + Age +
Female + Married + Wealth_level + (1 | Country)
Data: ds
REML criterion at convergence: 2526
Scaled residuals:
Min 1Q Median 3Q Max
-5.087 -0.593 0.033 0.677 4.070
Random effects:
Groups Name Variance Std.Dev.
Country (Intercept) 0.0192 0.139
Residual 0.3797 0.616
Number of obs: 1315, groups: Country, 4
Fixed effects:
Estimate Std. Error df t value
(Intercept) 2.791389 0.186736 126.440741 14.95
IG_Fusion 0.076776 0.022317 1301.703278 3.44
IG_Identification 0.124264 0.024302 1299.630921 5.11
OG_Bonds 0.017347 0.011905 1301.910330 1.46
Empathic_concern 0.038961 0.018456 1301.259994 2.11
Perspective_taking 0.174826 0.020026 1301.014809 8.73
Age -0.000787 0.001384 1301.732453 -0.57
Female1 0.039063 0.035127 1301.240598 1.11
MarriedOther -0.015903 0.081643 1300.212099 -0.19
MarriedUnmarried 0.088790 0.040799 1300.996291 2.18
Wealth_level2 0.164497 0.044162 1201.020084 3.72
Wealth_level3 0.138814 0.063917 1244.155630 2.17
Wealth_level4 -0.035545 0.130529 1299.892380 -0.27
Pr(>|t|)
(Intercept) < 0.0000000000000002 ***
IG_Fusion 0.0006 ***
IG_Identification 0.00000036 ***
OG_Bonds 0.1453
Empathic_concern 0.0350 *
Perspective_taking < 0.0000000000000002 ***
Age 0.5694
Female1 0.2663
MarriedOther 0.8456
MarriedUnmarried 0.0297 *
Wealth_level2 0.0002 ***
Wealth_level3 0.0301 *
Wealth_level4 0.7854
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Display code
tab_model(m01)
Perception_religious_freedom
Predictors
Estimates
CI
p
(Intercept)
2.79
2.43 – 3.16
<0.001
IG Fusion
0.08
0.03 – 0.12
0.001
IG Identification
0.12
0.08 – 0.17
<0.001
OG Bonds
0.02
-0.01 – 0.04
0.145
Empathic concern
0.04
0.00 – 0.08
0.035
Perspective taking
0.17
0.14 – 0.21
<0.001
Age
-0.00
-0.00 – 0.00
0.569
Female [1]
0.04
-0.03 – 0.11
0.266
Married [Other]
-0.02
-0.18 – 0.14
0.846
Married [Unmarried]
0.09
0.01 – 0.17
0.030
Wealth level [2]
0.16
0.08 – 0.25
<0.001
Wealth level [3]
0.14
0.01 – 0.26
0.030
Wealth level [4]
-0.04
-0.29 – 0.22
0.785
Random Effects
σ2
0.38
τ00Country
0.02
ICC
0.05
N Country
4
Observations
1315
Marginal R2 / Conditional R2
0.170 / 0.210
Here, marginal R sq is much higher compared to previous model. Adding individual-level predictors significantly increases explanatory power of the model. Again, evidence that most of the variation is at individual-level differences.
Display code
## Change class of all models so we can use stargazer():class(m00) <-"lmerMod"class(m01) <-"lmerMod"## Tabulated results:stargazer(m00, m01,type ="html", star.cutoffs =c(0.05, 0.01, 0.001),out ="table1.html")
Display code
htmltools::includeHTML("table1.html")
Dependent variable:
SPRF
Perception_religious_freedom
(1)
(2)
IG_Fusion
0.077***
(0.022)
IG_Identification
0.120***
(0.024)
OG_Bonds
0.017
(0.012)
Empathic_concern
0.039*
(0.018)
Perspective_taking
0.170***
(0.020)
Age
-0.001
(0.001)
Female1
0.039
(0.035)
MarriedOther
-0.016
(0.082)
MarriedUnmarried
0.089*
(0.041)
Wealth_level2
0.160***
(0.044)
Wealth_level3
0.140*
(0.064)
Wealth_level4
-0.036
(0.130)
Constant
5.400***
2.800***
(0.082)
(0.190)
Observations
1,497
1,315
Log Likelihood
-1,534.000
-1,263.000
Akaike Inf. Crit.
3,074.000
2,556.000
Bayesian Inf. Crit.
3,090.000
2,634.000
Note:
*p<0.05; **p<0.01; ***p<0.001
Random intercept models: Imagistic predictors
Display code
## Varying intercept models with individual-level predictors:m02 <-lmer(Perception_religious_freedom~Event_shared_perception+Event_episodic_recall+ Event_reflection+Event_positive_affect+Event_negative_affect+ Event_transformative_indiv+Event_transformative_group+ Age+Female+Married+Wealth_level+ (1| Country), data = ds)summary(m02)
Here, marginal R sq is much higher compared to previous model. Adding individual-level predictors significantly increases explanatory power of the model. Again, evidence that most of the variation is at individual-level differences.
Display code
## Change class of all models so we can use stargazer():class(m00) <-"lmerMod"class(m01) <-"lmerMod"class(m02) <-"lmerMod"## Tabulated results:stargazer(m00, m01, m02,type ="html", star.cutoffs =c(0.05, 0.01, 0.001),out ="table1.html")
Display code
htmltools::includeHTML("table1.html")
Dependent variable:
SPRF
Perception_religious_freedom
(1)
(2)
(3)
IG_Fusion
0.077***
(0.022)
IG_Identification
0.120***
(0.024)
OG_Bonds
0.017
(0.012)
Empathic_concern
0.039*
(0.018)
Perspective_taking
0.170***
(0.020)
Event_shared_perception
0.018
(0.015)
Event_episodic_recall
0.150***
(0.018)
Event_reflection
0.007
(0.015)
Event_positive_affect
0.013
(0.010)
Event_negative_affect
0.034***
(0.009)
Event_transformative_indiv
0.045**
(0.015)
Event_transformative_group
0.037**
(0.014)
Age
-0.001
0.0001
(0.001)
(0.001)
Female1
0.039
0.026
(0.035)
(0.034)
MarriedOther
-0.016
-0.044
(0.082)
(0.082)
MarriedUnmarried
0.089*
0.083*
(0.041)
(0.040)
Wealth_level2
0.160***
0.160***
(0.044)
(0.044)
Wealth_level3
0.140*
0.018
(0.064)
(0.063)
Wealth_level4
-0.036
-0.170
(0.130)
(0.130)
Constant
5.400***
2.800***
3.700***
(0.082)
(0.190)
(0.170)
Observations
1,497
1,315
1,369
Log Likelihood
-1,534.000
-1,263.000
-1,320.000
Akaike Inf. Crit.
3,074.000
2,556.000
2,673.000
Bayesian Inf. Crit.
3,090.000
2,634.000
2,762.000
Note:
*p<0.05; **p<0.01; ***p<0.001
Histogram: Perception of Religious Freedom
Display code
summary(ds$SPRF)
Min. 1st Qu. Median Mean 3rd Qu. Max. NA's
2 5 6 5 6 7 45
Section 3. Outcome: Experience of Religious Freedom
Unconditional means model
Also called “varying intercept model with no predictors” (Gelman and Hill, 2016, Chapter 12). Allows intercepts to randomly vary across countries:
Display code
ds$Exp_religious_freedom <- ds$Exp_religious_freedom## Varying intercept model with no predictors:m10<-lmer(Exp_religious_freedom ~1+ (1| Country), data = ds)summary(m10)
Linear mixed model fit by REML. t-tests use Satterthwaite's method [
lmerModLmerTest]
Formula: Exp_religious_freedom ~ 1 + (1 | Country)
Data: ds
REML criterion at convergence: 5077
Scaled residuals:
Min 1Q Median 3Q Max
-3.924 -0.515 0.369 0.705 1.210
Random effects:
Groups Name Variance Std.Dev.
Country (Intercept) 0.129 0.36
Residual 1.743 1.32
Number of obs: 1492, groups: Country, 4
Fixed effects:
Estimate Std. Error df t value Pr(>|t|)
(Intercept) 5.841 0.183 3.037 31.9 0.000062 ***
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Random effects:
Variance for Intercept = 0.141. This is the variance of the means across level 1 categories (countries). Residual = 1.7. Variance explained by level 1 residuals (everything that’s not in level 1).
Display code
tab_model(m10)
Exp_religious_freedom
Predictors
Estimates
CI
p
(Intercept)
5.84
5.48 – 6.20
<0.001
Random Effects
σ2
1.74
τ00Country
0.13
ICC
0.07
N Country
4
Observations
1492
Marginal R2 / Conditional R2
0.000 / 0.069
We can see that ICC = 0.08. Lower ICC = low variance explained across groups. In this case, most of the variability is at individual-level (not group level). There is no significantly different patterns between countries.
Random intercept models
Also called “varying intercept model with individual-level predictors” (Gelman and Hill, 2016, Chapter 12).
Display code
## Varying intercept models with individual-level predictors:m11 <-lmer(Exp_religious_freedom~IG_Fusion+IG_Identification+OG_Bonds+Empathic_concern+ Perspective_taking+Age+Female+Married+Wealth_level+ (1| Country), data = ds)summary(m11)
Linear mixed model fit by REML. t-tests use Satterthwaite's method [
lmerModLmerTest]
Formula: Exp_religious_freedom ~ IG_Fusion + IG_Identification + OG_Bonds +
Empathic_concern + Perspective_taking + Age + Female + Married +
Wealth_level + (1 | Country)
Data: ds
REML criterion at convergence: 4251
Scaled residuals:
Min 1Q Median 3Q Max
-4.592 -0.524 0.233 0.696 2.405
Random effects:
Groups Name Variance Std.Dev.
Country (Intercept) 0.0852 0.292
Residual 1.4240 1.193
Number of obs: 1316, groups: Country, 4
Fixed effects:
Estimate Std. Error df t value
(Intercept) 4.50051 0.36592 98.91990 12.30
IG_Fusion 0.04887 0.04321 1302.16380 1.13
IG_Identification 0.11225 0.04709 1300.48983 2.38
OG_Bonds -0.20987 0.02305 1302.97689 -9.10
Empathic_concern 0.17408 0.03574 1302.85029 4.87
Perspective_taking -0.00743 0.03884 1302.50507 -0.19
Age 0.00260 0.00267 1302.57372 0.97
Female1 0.22350 0.06807 1302.67705 3.28
MarriedOther 0.22189 0.15809 1302.28377 1.40
MarriedUnmarried 0.04714 0.07906 1301.68725 0.60
Wealth_level2 -0.03748 0.08541 1235.25598 -0.44
Wealth_level3 -0.34497 0.12377 1264.82102 -2.79
Wealth_level4 -0.49288 0.25279 1302.08011 -1.95
Pr(>|t|)
(Intercept) < 0.0000000000000002 ***
IG_Fusion 0.2583
IG_Identification 0.0173 *
OG_Bonds < 0.0000000000000002 ***
Empathic_concern 0.0000013 ***
Perspective_taking 0.8484
Age 0.3303
Female1 0.0011 **
MarriedOther 0.1607
MarriedUnmarried 0.5512
Wealth_level2 0.6609
Wealth_level3 0.0054 **
Wealth_level4 0.0514 .
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Display code
tab_model(m11)
Exp_religious_freedom
Predictors
Estimates
CI
p
(Intercept)
4.50
3.78 – 5.22
<0.001
IG Fusion
0.05
-0.04 – 0.13
0.258
IG Identification
0.11
0.02 – 0.20
0.017
OG Bonds
-0.21
-0.26 – -0.16
<0.001
Empathic concern
0.17
0.10 – 0.24
<0.001
Perspective taking
-0.01
-0.08 – 0.07
0.848
Age
0.00
-0.00 – 0.01
0.330
Female [1]
0.22
0.09 – 0.36
0.001
Married [Other]
0.22
-0.09 – 0.53
0.161
Married [Unmarried]
0.05
-0.11 – 0.20
0.551
Wealth level [2]
-0.04
-0.21 – 0.13
0.661
Wealth level [3]
-0.34
-0.59 – -0.10
0.005
Wealth level [4]
-0.49
-0.99 – 0.00
0.051
Random Effects
σ2
1.42
τ00Country
0.09
ICC
0.06
N Country
4
Observations
1316
Marginal R2 / Conditional R2
0.122 / 0.172
Here, marginal R sq is much higher compared to previous model. Adding individual-level predictors significantly increases explanatory power of the model. Again, evidence that most of the variation is at individual-level differences.
Display code
## Change class of all models so we can use stargazer():class(m10) <-"lmerMod"class(m11) <-"lmerMod"## Tabulated results:stargazer(m10, m11,type ="html", star.cutoffs =c(0.05, 0.01, 0.001),out ="table1.html")
Display code
htmltools::includeHTML("table1.html")
Dependent variable:
Exp_religious_freedom
(1)
(2)
IG_Fusion
0.049
(0.043)
IG_Identification
0.110*
(0.047)
OG_Bonds
-0.210***
(0.023)
Empathic_concern
0.170***
(0.036)
Perspective_taking
-0.007
(0.039)
Age
0.003
(0.003)
Female1
0.220**
(0.068)
MarriedOther
0.220
(0.160)
MarriedUnmarried
0.047
(0.079)
Wealth_level2
-0.037
(0.085)
Wealth_level3
-0.340**
(0.120)
Wealth_level4
-0.490
(0.250)
Constant
5.800***
4.500***
(0.180)
(0.370)
Observations
1,492
1,316
Log Likelihood
-2,538.000
-2,125.000
Akaike Inf. Crit.
5,083.000
4,281.000
Bayesian Inf. Crit.
5,099.000
4,359.000
Note:
*p<0.05; **p<0.01; ***p<0.001
Random intercept models: Imagistic predictors
Display code
## Varying intercept models with individual-level predictors:m12 <-lmer(Exp_religious_freedom~Event_shared_perception+Event_episodic_recall+ Event_reflection+Event_positive_affect+Event_negative_affect+ Event_transformative_indiv+Event_transformative_group+ Age+Female+Married+Wealth_level+ (1| Country), data = ds)summary(m12)
Here, marginal R sq is much higher compared to previous model. Adding individual-level predictors significantly increases explanatory power of the model. Again, evidence that most of the variation is at individual-level differences.
Display code
## Change class of all models so we can use stargazer():class(m10) <-"lmerMod"class(m11) <-"lmerMod"class(m12) <-"lmerMod"## Tabulated results:stargazer(m10, m11, m12,type ="html", star.cutoffs =c(0.05, 0.01, 0.001),out ="table1.html")